[BUGFIX] Fix monitoring of mounted pages

EXT:solr fails to detect changes on mounted pages correctly in some special occasions: * mounted pages not considered during garbage removal / ce updates * no_search_sub_entries flag not respected by GarbageRemover strategies * existing queue items of mounted pages not updated By adjusting the monitoring related components the issues are fixed, affected components: * GarbageRemover strategies * page initializer * RecordMonitor * DataUpdateHandler Fixes: TYPO3-Solr#4275
dkd-friedrich · Jan 7, 2025 · b43c563 · b43c563
1 parent fe71a4f
commit b43c563
Show file tree

Hide file tree

Showing 11 changed files with 470 additions and 28 deletions.
diff --git a/Classes/Domain/Index/Queue/GarbageRemover/AbstractStrategy.php b/Classes/Domain/Index/Queue/GarbageRemover/AbstractStrategy.php
@@ -23,7 +23,9 @@
 use ApacheSolrForTypo3\Solr\IndexQueue\Queue;
 use ApacheSolrForTypo3\Solr\IndexQueue\QueueInterface;
 use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager;
+use ApacheSolrForTypo3\Solr\Traits\SkipRecordByRootlineConfigurationTrait;
 use Doctrine\DBAL\Exception as DBALException;
+use TYPO3\CMS\Backend\Utility\BackendUtility;
 use TYPO3\CMS\Core\Utility\GeneralUtility;
 use UnexpectedValueException;
 
@@ -33,6 +35,8 @@
  */
 abstract class AbstractStrategy
 {
+    use SkipRecordByRootlineConfigurationTrait;
+
     protected QueueInterface $queue;
     protected ConnectionManager $connectionManager;
     protected SiteRepository $siteRepository;
@@ -82,7 +86,15 @@ protected function deleteInSolrAndRemoveFromIndexQueue(string $table, int $uid):
     protected function deleteInSolrAndUpdateIndexQueue(string $table, int $uid): void
     {
         $this->deleteIndexDocuments($table, $uid);
-        $this->queue->updateItem($table, $uid);
+
+        $record = BackendUtility::getRecord($table, $uid);
+        if ($record === null) {
+            return;
+        }
+
+        if (!$this->skipRecordByRootlineConfiguration((int)($record['pid'] ?? 0))) {
+            $this->queue->updateItem($table, $uid);
+        }
     }
 
     /**

diff --git a/Classes/Domain/Index/Queue/GarbageRemover/PageStrategy.php b/Classes/Domain/Index/Queue/GarbageRemover/PageStrategy.php
@@ -15,9 +15,11 @@
 
 namespace ApacheSolrForTypo3\Solr\Domain\Index\Queue\GarbageRemover;
 
+use ApacheSolrForTypo3\Solr\Domain\Index\Queue\RecordMonitor\Helper\MountPagesUpdater;
 use ApacheSolrForTypo3\Solr\Domain\Site\Exception\UnexpectedTYPO3SiteInitializationException;
 use Doctrine\DBAL\Exception as DBALException;
 use TYPO3\CMS\Backend\Utility\BackendUtility;
+use TYPO3\CMS\Core\Utility\GeneralUtility;
 
 /**
  * Class PageStrategy
@@ -42,6 +44,23 @@ protected function removeGarbageOfByStrategy(string $table, int $uid): void
         }
     }
 
+    /**
+     * Deletes a page from Solr and updates the item in the index queue
+     * The MountPagesUpdater takes care of mounted pages
+     *
+     * @throws DBALException
+     * @throws UnexpectedTYPO3SiteInitializationException
+     */
+    protected function deleteInSolrAndUpdateIndexQueue(string $table, int $uid): void
+    {
+        parent::deleteInSolrAndUpdateIndexQueue($table, $uid);
+
+        if ($table === 'pages') {
+            $mountPagesUpdater = GeneralUtility::makeInstance(MountPagesUpdater::class);
+            $mountPagesUpdater->update($uid);
+        }
+    }
+
     /**
      * Determines the relevant page id for a content element update. Deletes the page from solr and requeues the
      * page for a reindex.

diff --git a/Classes/Domain/Index/Queue/UpdateHandler/DataUpdateHandler.php b/Classes/Domain/Index/Queue/UpdateHandler/DataUpdateHandler.php
@@ -30,6 +30,7 @@
 use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager;
 use ApacheSolrForTypo3\Solr\System\Records\Pages\PagesRepository;
 use ApacheSolrForTypo3\Solr\System\TCA\TCAService;
+use ApacheSolrForTypo3\Solr\Traits\SkipRecordByRootlineConfigurationTrait;
 use ApacheSolrForTypo3\Solr\Util;
 use Doctrine\DBAL\Exception as DBALException;
 use Throwable;
@@ -46,6 +47,8 @@
  */
 class DataUpdateHandler extends AbstractUpdateHandler
 {
+    use SkipRecordByRootlineConfigurationTrait;
+
     /**
      * List of fields in the update field array that
      * are required for processing
@@ -388,7 +391,9 @@ protected function processPageRecord(int $uid, int $pid, array $updatedFields =
         }
         $rootPageIds = [$configurationPageId];
 
-        $this->processRecord('pages', $uid, $rootPageIds);
+        if (!$this->skipRecordByRootlineConfiguration($uid)) {
+            $this->processRecord('pages', $uid, $rootPageIds);
+        }
 
         $this->updateCanonicalPages($uid);
         $this->mountPageUpdater->update($uid);

diff --git a/Classes/IndexQueue/Initializer/Page.php b/Classes/IndexQueue/Initializer/Page.php
@@ -23,6 +23,7 @@
 use PDO;
 use Throwable;
 use TYPO3\CMS\Backend\Utility\BackendUtility;
+use TYPO3\CMS\Core\Context\Context;
 use TYPO3\CMS\Core\Database\Connection;
 use TYPO3\CMS\Core\Database\ConnectionPool;
 use TYPO3\CMS\Core\Messaging\FlashMessage;
@@ -224,8 +225,22 @@ protected function addMountedPagesToIndexQueue(array $mountedPages, array $mount
     {
         $mountPointIdentifier = $this->getMountPointIdentifier($mountProperties);
         $mountPointPageIsWithExistingQueueEntry = $this->queueItemRepository->findPageIdsOfExistingMountPagesByMountIdentifier($mountPointIdentifier);
-        $mountedPagesThatNeedToBeAdded = array_diff($mountedPages, $mountPointPageIsWithExistingQueueEntry);
 
+        // update existing queue entries to trigger reindexing
+        array_walk(
+            $mountPointPageIsWithExistingQueueEntry,
+            function (int $pageUid): void {
+                $tstamp = GeneralUtility::makeInstance(Context::class)
+                    ->getPropertyFromAspect('date', 'timestamp') ?? time();
+                $items = $this->queueItemRepository->findItemsByItemTypeAndItemUid('pages', $pageUid);
+                foreach ($items as $item) {
+                    $this->queueItemRepository->updateChangedTimeByItem($item, $tstamp);
+                }
+            }
+        );
+
+        // add new items if necessary
+        $mountedPagesThatNeedToBeAdded = array_diff($mountedPages, $mountPointPageIsWithExistingQueueEntry);
         if (count($mountedPagesThatNeedToBeAdded) === 0) {
             //nothing to do
             return;

diff --git a/Classes/IndexQueue/RecordMonitor.php b/Classes/IndexQueue/RecordMonitor.php
@@ -22,14 +22,13 @@
 use ApacheSolrForTypo3\Solr\Domain\Index\Queue\UpdateHandler\Events\RecordUpdatedEvent;
 use ApacheSolrForTypo3\Solr\Domain\Index\Queue\UpdateHandler\Events\VersionSwappedEvent;
 use ApacheSolrForTypo3\Solr\Traits\SkipMonitoringTrait;
+use ApacheSolrForTypo3\Solr\Traits\SkipRecordByRootlineConfigurationTrait;
 use ApacheSolrForTypo3\Solr\Util;
 use Psr\EventDispatcher\EventDispatcherInterface;
 use TYPO3\CMS\Backend\Utility\BackendUtility;
 use TYPO3\CMS\Core\DataHandling\DataHandler;
-use TYPO3\CMS\Core\Exception\Page\PageNotFoundException;
 use TYPO3\CMS\Core\Utility\GeneralUtility;
 use TYPO3\CMS\Core\Utility\MathUtility;
-use TYPO3\CMS\Core\Utility\RootlineUtility;
 
 /**
  * A class that monitors changes to the records so that the changed record get
@@ -38,6 +37,7 @@
 class RecordMonitor
 {
     use SkipMonitoringTrait;
+    use SkipRecordByRootlineConfigurationTrait;
 
     protected array $trackedRecords = [];
     protected EventDispatcherInterface $eventDispatcher;
@@ -144,7 +144,11 @@ public function processDatamap_afterDatabaseOperations(
                 $recordPid = $recordInfo['pid'] ?? null;
             }
         }
-        if (!is_null($recordPid) && $this->skipRecordByRootlineConfiguration((int)$recordPid)) {
+        if (
+            !is_null($recordPid)
+            && !GeneralUtility::inList('pages,tt_content', $table)
+            && $this->skipRecordByRootlineConfiguration((int)$recordPid)
+        ) {
             return;
         }
 
@@ -164,24 +168,4 @@ public function processDatamap_afterDatabaseOperations(
             new RecordUpdatedEvent((int)$recordUid, $table, $fields)
         );
     }
-
-    /**
-     * Check if at least one page in the record's rootline is configured to exclude sub-entries from indexing
-     */
-    protected function skipRecordByRootlineConfiguration(int $pid): bool
-    {
-        /** @var RootlineUtility $rootlineUtility */
-        $rootlineUtility = GeneralUtility::makeInstance(RootlineUtility::class, $pid);
-        try {
-            $rootline = $rootlineUtility->get();
-        } catch (PageNotFoundException) {
-            return true;
-        }
-        foreach ($rootline as $page) {
-            if (isset($page['no_search_sub_entries']) && $page['no_search_sub_entries']) {
-                return true;
-            }
-        }
-        return false;
-    }
 }
diff --git a/Classes/Traits/SkipRecordByRootlineConfigurationTrait.php b/Classes/Traits/SkipRecordByRootlineConfigurationTrait.php
@@ -0,0 +1,45 @@
+<?php
+
+declare(strict_types=1);
+
+/*
+ * This file is part of the TYPO3 CMS project.
+ *
+ * It is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU General Public License, either version 2
+ * of the License, or any later version.
+ *
+ * For the full copyright and license information, please read the
+ * LICENSE.txt file that was distributed with this source code.
+ *
+ * The TYPO3 project - inspiring people to share!
+ */
+
+namespace ApacheSolrForTypo3\Solr\Traits;
+
+use TYPO3\CMS\Core\Exception\Page\PageNotFoundException;
+use TYPO3\CMS\Core\Utility\GeneralUtility;
+use TYPO3\CMS\Core\Utility\RootlineUtility;
+
+trait SkipRecordByRootlineConfigurationTrait
+{
+    /**
+     * Check if at least one page in the record's rootline is configured to exclude sub-entries from indexing
+     */
+    protected function skipRecordByRootlineConfiguration(int $pid): bool
+    {
+        /** @var RootlineUtility $rootlineUtility */
+        $rootlineUtility = GeneralUtility::makeInstance(RootlineUtility::class, $pid);
+        try {
+            $rootline = $rootlineUtility->get();
+        } catch (PageNotFoundException) {
+            return true;
+        }
+        foreach ($rootline as $page) {
+            if (isset($page['no_search_sub_entries']) && $page['no_search_sub_entries']) {
+                return true;
+            }
+        }
+        return false;
+    }
+}
diff --git a/Tests/Integration/IndexQueue/Fixtures/mount_page_updates.csv b/Tests/Integration/IndexQueue/Fixtures/mount_page_updates.csv
@@ -0,0 +1,25 @@
+"pages",
+,"uid","pid","tstamp","hidden","doktype","mount_pid","mount_pid_ol","slug","title","no_search_sub_entries"
+,2,1,1736145036,0,254,0,0,"/sysfolder","Sys folder (no_search_sub_entries)",1
+,20,2,1736145036,0,1,0,0,"/mounted-page-in-sysfolder","Mounted page in sysfolder",0
+,21,1,1736145036,0,1,0,0,"/mounted-page-on-root","Mounted page on root page",0
+,22,2,1736145036,0,1,0,0,"/another-mounted-page-in-sysfolder","Another mounted page in sysfolder",0
+,30,1,1736145036,0,7,20,1,"/mount-point-1","Mount point for page in sysfolder",0
+,31,1,1736145036,0,7,21,1,"/mount-point-2","Mount point for page on root page",0
+,32,1,1736145036,0,7,22,1,"/mount-point-3","Mount point for another page in sysfolder",0
+"tt_content",
+,"uid","pid","header","hidden"
+,20,20,"ce on mounted page",0
+,21,20,"hidden ce on mounted page",1
+,22,21,"ce on mounted page",0
+,23,21,"hidden ce on mounted page",1
+,24,22,"ce on another mounted page",0
+,25,22,"hidden ce on another mounted page",1
+"tx_solr_indexqueue_item"
+,"uid","root","item_type","item_uid","indexing_configuration","has_indexing_properties","pages_mountidentifier","errors"
+,1,1,"pages",22,"pages",1,"22-32-1",""
+"tx_solr_indexqueue_indexing_property"
+,"uid","root","item_id","property_key","property_value"
+,1,1,1,"mountPageSource",22
+,2,1,1,"mountPageDestination",32
+,3,1,1,"isMountedPage",1